INDEX

Explanations

references to points or scores in competitive contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-fields

-0.16

ugo

-0.16

 soccer

-0.15

fields

-0.15

kan

-0.15

è¶³çĲĥ

-0.15

 FIFA

-0.14

 ì¶ķêµ¬

-0.14

 Wass

-0.14

 fields

-0.14

POSITIVE LOGITS

NI

0.23

 Devils

0.22

ice

0.21

Ice

0.21

GB

0.20

 Freeze

0.20

Ice

0.20

 Guild

0.20

Sol

0.19

NI

0.19

Activations Density 0.000%

references to points or scores in competitive contexts

No Comments

No Known Activations